Try with an electrolytic and then if you have something better try again.
Finally you will have your own opinion.
Yes, that seems about right. The P3A is about 60W/ch, which means you need at least about 6000uF/rail for each channel, rule of thumb. 10kuF/rail would be a good figure. So, 40kuF for a stereo box.
Some audiophiles for some reason like one bridge between ground and +ve rail, a second bridge between ground and -ve rail. I have tried tracing the reasoning behind this strange practice through various discussions, studying what much more mature designers than me feel. I have not found any veteran of power amp design supporting the two-bridge approach. Some have questioned why anyone would use two bridges, because it adds diode drops in the path of the ground rail, which is extremely important to keep clean. Any voltage drop in the path of (often very high) current in the ground rail can harm amplifier performance. This is not my opinion -- I've seen others say this (can't point to quotes right now, sorry).
FWIW, I too use a single bridge per amp. No diodes on the ground rail for me.
I am fairly certain there will be vocal supporters of the two-bridge approach here, so take your pick.
By the way, about rail filtering capacitors -- I use a "pi filter" in the supply rails. Helps clean up HF noise and diode switching noise from the rails, and this improves the performance of power amps which have less than stellar PSRR. The P3A is a relatively simple topology and has not-quite-top-notch PSRR.
A pi filter is made of two capacitor banks separated by a small resistor. I use a 10W or 20W half-Ohm resistor. Imagine that you are replacing one capacitor with a C-R-C structure, where the first C is the original shunt capacitor, then a series resistor on the rail, then a second shunt capacitor of the same value. Where conventional designs would use a 10KuF capacitor, I will use a 10KuF, followed by a half-Ohm resistor, followed by a second 10KuF capacitor.
This, if done, must be done for each rail. So, the total capacitors I would use for a stereo P3A would go from 40KuF to 80KuF.
A pi filter is made of two capacitor banks separated by a small resistor. I use a 10W or 20W half-Ohm resistor. Imagine that you are replacing one capacitor with a C-R-C structure, where the first C is the original shunt capacitor, then a series resistor on the rail, then a second shunt capacitor of the same value. Where conventional designs would use a 10KuF capacitor, I will use a 10KuF, followed by a half-Ohm resistor, followed by a second 10KuF capacitor.
This, if done, must be done for each rail. So, the total capacitors I would use for a stereo P3A would go from 40KuF to 80KuF.
many thanks.
I've seen double rectifiers on Pass amps F5 and F4, Penasa's MY_REF, and also on the italian site TNT audio. This said, I also read that double bridging helps lets the initial current on powering on to be splitted between the two bridges--so less stress on diodes.
I also read about the diode drop in the path of the ground rail.
I think I'll use a double rectifier if that will not hurt anyway...
I've seen double rectifiers on Pass amps F5 and F4, Penasa's MY_REF, and also on the italian site TNT audio. This said, I also read that double bridging helps lets the initial current on powering on to be splitted between the two bridges--so less stress on diodes.
I also read about the diode drop in the path of the ground rail.
I think I'll use a double rectifier if that will not hurt anyway...
using the dual secondary with a dual bridge rectifier gives TWO ISOLATED supplies.
These isolated supplies can be connection in series. This junction is usually referred to as "ground". But it is NOT a "ground", it is more correctly referred to as Power Zero Volts.
The amplifier gets connected to this Power Zero Volts. There is no diode between amplifier and Power Zero Volts.
Or in your language there is no diode in the "ground"
you have not searched fully. There are many here and elsewhere that claim the dual secondary to dual rectifier is the better way to avoid asymetrically loading the transformer.
Not true. The sudden current requirements in a Class B amp are supposed to be met by the downstream capacitor bank of the Pi filter. Without a Pi filter, they are supposed to be met by the capacitor bank anyway. If the capacitor bank in a normal PSU is getting discharged anyway, then the ripple will pass through and you have bigger problems than just high source impedance.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.